Validation of a wireless high-resolution nighttime oximeter for the diagnosis of obstructive sleep apnea at home

Our study showed that ODM using high-resolution wireless oximeter and cloud-based automated algorithm for desaturation detection (Biologix) is an accurate method for diagnosis of OSA and determination of OSA severity at home in patients with suspected OSA. Evidence derived from observation of good performance in detecting at least mild, moderate and severe OSA, as evidenced by high AUC determined by the ROC curve (Fig. 4), high sensitivity, specificity, PPV, VPN, LR+, and LM− (Table 3). Finally, the variability between HST-AHI and ODM-ODI was lower than that between HST-AHI variability from night to night (Fig. 3).

PSG is considered the gold standard for diagnosing OSA17. However, the HST has been widely validated for the diagnosis and management of OSA15,17,18. Despite the planned move to improve diagnostic accessibility, there remains an unacceptable level of OSA under diagnosis10.15. Nighttime oximetry has long been advocated as an attractive technique for diagnosing OSA. However, a recent meta-analysis has shown that the performance of oximeters designed for the diagnosis of OSA is highly variable and depends on study design and technology. The meta-analysis found eight studies comparing PSG to ODM. However, only three of them performed simultaneous monitoring of PSG and ODM. The authors concluded that nocturnal oximetry was adequate to detect moderate to severe OSA when the 4% desaturation criterion was used.16. We previously showed good performance of ODM-Biologix compared to PSG-AHI Sleep Lab14. Our study extends these results by showing that the ODM-Biologix has good accuracy for home diagnosis for all severity ranges, including at least mild, moderate-to-severe, and severe OSAS (Table 3). Contrary to our main hypothesis that night-to-night variability of HST would be similar to HST-AHI versus ODM-ODI variability, we found significantly lower variability of HST versus ODM than variability night-night of the HST-AHI (Fig. 3). The relatively high overnight variability of HST-AHI in our study may be partially explained by the small sample size. However, a recent large sample of patients who repeated HST for three nights (10,340) showed that the confidence interval of HST-IAH from one night to the next was approximately 21 events/h.16, which is similar to the variability between HST-AHI and ODM-ODI found in our study (25.5 events/h). Similar levels of night-to-night variability were also seen in the PSG sleep lab.19. Although in clinical practice treatment decisions are made on the basis of a single sleep study, the well-known variability from night to night indicates that repeat sleep studies should be useful, particularly when there is a mismatch between clinical symptoms and sleep study results. However, this hypothesis has not been widely explored due to the inherent difficulties of repeating sleep studies. Therefore, simple, easy-to-use, and accurate techniques for OSA diagnosis that can be easily repeated, such as Biologix, can help improve clinical decisions.

Biologix has several potential benefits that can help address the challenge of OSA underdiagnosis. The cost of Biologix is ​​an order of magnitude lower than that of HST equipment (less than $100 versus a few thousand dollars, respectively). HST studies have costs associated with disposable materials, including batteries and nasal cannulas, which are absent from the Biologix methodology. The HST requires specialized sleep centers and personnel to analyze and interpret the HST study. On the other hand, the Biologix platform automatically analyzed the data. Taken together, Biologix enables a paradigm shift that facilitates direct access to OSA diagnosis by non-sleep related specialist professionals. This methodology can help apply sleep apnea diagnosis and treatment programs directly in primary care centers20.21.

Although we considered the HST-AHI as the gold standard in this study, the concept that AHI is the most important parameter in defining the severity of OSA has been questioned.22. For example, there is growing evidence that oximeter-derived parameters, such as ODI and hypoxic load, are more relevant than AHI in determining the association with hypertension and in predicting the future cardiovascular events.23. It should also be emphasized that the treatment decision depends on the patient’s symptoms and the presence of comorbidities. Finally, the treatment decision will always depend on the experience of the center as well as the willingness of the patient to accept specific treatment modalities, such as surgery or CPAP. Therefore, the current evidence of a good correlation between HST-AHI and ODM-ODI is strong evidence that the Biologix platform provides enough information for clinical treatment decisions in patients with suspected OSA. Additionally, the fact that Biologix is ​​inexpensive, minimally invasive and accurate, leads to the possibility of multiple studies. Biologix can address the intrinsic uncertainty derived from overnight variability, which is largely ignored in clinical practice due to the economic and logistical challenges of conducting multiple sleep studies. Repeated sleep studies can also address fluctuations in OSA severity associated with alcohol consumption. The device may also be particularly useful for monitoring patients under different treatments, such as during titration of mandibular advancement devices, weight loss and alternative treatments.

Our study had several limitations. First, Biologix was not offered to all patients who participated in this accelerated OSAS diagnosis due to logistical limitations. However, we studied a large sample of patients with suspected OSA (Table 1). Second, ODM depends on smartphones with internet connections, which may not be readily available in some regions. Third, the number of patients using HST, the variability from night to night, was lower than that of patients using Biologix and HST. However, our variability in OSA severity from night to night is consistent with previous studies19.24. Fourth, in this study, ODM was compared to HST which is considered the gold standard. However, the HST has potential limitations as it detects sleep, sleep stages, and sleep arousals which may be relevant parameters in OSA patient subgroups. Finally, ODM does not distinguish between central and obstructive events and has not been validated in patients with significant comorbidities, such as heart failure and severe lung disease.

In conclusion, ODM-Biologix is ​​a simple and effective method for the diagnosis of OSAS at home in patients suspected of OSAS. As technology advances, new avenues need to be explored, such as the inclusion of more sophisticated signal analysis of pulse rate and sleep architecture using machine learning algorithms .